Keyword Cluster Architect Agent Skill

Keyword Cluster Architect

Map 50-200+ keywords into topical clusters grouped by search intent. Build a content roadmap for dominating a topic with hub-and-spoke architecture. Google rewards topical authority — this skill builds the strategic map that tells you exactly what content to create and in what order.

Stage

S3: Blog & SEO — This is the strategic planning layer FOR blog content. Before writing individual posts, you need a map of the entire keyword landscape organized into clusters.

When to Use

• User wants to plan SEO content strategy for a niche
• User asks about keyword research, clustering, or topical authority
• User says "keyword", "SEO plan", "content roadmap", "topic cluster", "hub and spoke"
• Before running affiliate-blog-builder — to know WHICH articles to write
• After monopoly-niche-finder — to map the keyword universe for the winning niche

Input Schema

niche: string                 # REQUIRED — the topic to cluster
                              # e.g., "AI video tools", "email marketing for SaaS"

seed_keywords: string[]       # OPTIONAL — starting keywords to expand from
                              # Default: auto-generated from niche

depth: string                 # OPTIONAL — "quick" (50 keywords) | "standard" (100) | "deep" (200+)
                              # Default: "standard"

affiliate_products: string[]  # OPTIONAL — products you promote (to prioritize commercial keywords)
                              # Default: none

Chaining from S1 monopoly-niche-finder: Use monopoly_niche.intersection as the niche input.

Workflow

Step 1: Generate Seed Keywords

If not provided, generate 5-10 seed keywords from the niche:

• Product-focused: "[product] review", "best [category]"
• Problem-focused: "how to [solve problem]", "[problem] solution"
• Comparison: "[product A] vs [product B]", "alternatives to [product]"
• Tutorial: "how to use [product]", "[product] tutorial"

Step 2: Expand Keywords

For each seed, use web_search to discover related keywords:

• Search: "[seed keyword]" — note related searches, People Also Ask
• Search: "[seed keyword] guide" OR "[seed keyword] tutorial" — informational variants
• Search: "best [seed keyword]" OR "[seed keyword] review" — commercial variants

Collect 50-200+ unique keywords depending on depth.

Step 3: Classify by Intent

Read shared/references/seo-strategy.md for clustering methodology.

Classify each keyword:

• Informational (I): Learning, how-to, what-is → blog posts, tutorials
• Commercial (C): Comparing, evaluating, reviewing → comparison posts, reviews
• Transactional (T): Ready to buy, pricing, discount → landing pages, deal pages
• Navigational (N): Brand-specific, login → skip (not your traffic to capture)

Step 4: Cluster by Topic

Group keywords that share the same search intent (would be answered by the same page):

Cluster: "[Main Topic]"
  Type: [I/C/T]
  Hub keyword: [highest volume keyword]
  Supporting keywords:
    - [keyword 1] — [est. volume]
    - [keyword 2] — [est. volume]
  Content type: [blog post / comparison / review / tutorial / landing page]
  Priority: [1-5 based on volume × intent × competition]

Step 5: Build Content Roadmap

Organize clusters into a hub-and-spoke map:

• Identify the hub page (broadest, highest-volume cluster)
• Connect spoke pages (specific clusters that link back to hub)
• Prioritize by: commercial intent first (revenue), then informational (traffic)
• Estimate effort: number of articles needed, suggested publishing cadence

Step 6: Self-Validation

• [ ] Clusters are based on actual search data, not guesses
• [ ] Each cluster has a clear search intent (I, C, or T)
• [ ] Hub-and-spoke structure is logical (hub is broad, spokes are specific)
• [ ] Priority ordering makes business sense (revenue-driving content first)
• [ ] Total content pieces are realistic for user's capacity

Output Schema

output_schema_version: "1.0.0"
keyword_clusters:
  niche: string
  total_keywords: number
  total_clusters: number

  hub:
    keyword: string
    cluster_name: string
    content_type: string
    priority: number

  clusters:
    - name: string
      intent: string          # "informational" | "commercial" | "transactional"
      hub_keyword: string
      keywords: string[]
      content_type: string    # "blog" | "comparison" | "review" | "tutorial" | "landing"
      priority: number        # 1-5
      estimated_volume: string

  content_roadmap:
    total_articles: number
    publishing_cadence: string
    priority_order: string[]  # Cluster names in order to write

  target_keywords: string[]   # Flat list of all keywords for chaining

chain_metadata:
  skill_slug: "keyword-cluster-architect"
  stage: "blog"
  timestamp: string
  suggested_next:
    - "affiliate-blog-builder"
    - "content-moat-calculator"
    - "comparison-post-writer"
    - "landing-page-creator"

Output Format

## Keyword Cluster Map: [Niche]

### Overview
- **Total keywords:** XXX
- **Clusters:** XX
- **Hub topic:** [main hub]
- **Content pieces needed:** XX articles

### Hub & Spoke Map

[HUB: Main Topic]

/ | | \

[Spoke] [Spoke] [Spoke] [Spoke]

| | | |

[Sub] [Sub] [Sub] [Sub]

### Clusters by Priority

#### Priority 1: [Cluster Name] (Commercial Intent)
- **Hub keyword:** [keyword] — [volume]
- **Content type:** [comparison / review]
- **Keywords:** [list]
- **Article idea:** [specific title]

#### Priority 2: [Cluster Name] (Informational Intent)
[same structure]

[Continue for all clusters]

### Content Roadmap
| Week | Cluster | Article | Intent | Priority |
|---|---|---|---|---|
| 1 | [cluster] | [title] | C | 1 |
| 2 | [cluster] | [title] | C | 1 |
| 3 | [cluster] | [title] | I | 2 |

### Next Steps
- Run `content-moat-calculator` to estimate effort for topical authority
- Run `affiliate-blog-builder` for Priority 1 articles
- Run `comparison-post-writer` for commercial clusters

Error Handling

• Niche too broad: "This niche is very broad. Let me narrow to a sub-niche for more actionable clusters. Or run monopoly-niche-finder first."
• No search volume: "This niche may be too narrow for significant search traffic. Consider broadening slightly."
• Too many keywords: Group aggressively into fewer clusters. Quality of clustering > quantity of keywords.
• No commercial intent keywords: Flag as concern — hard to monetize through affiliate without commercial intent. Suggest adjacent niches.

Examples

Example 1: "Map keywords for AI video tools"

→ Seeds: "best AI video tools", "AI video generator", "HeyGen review". Expand to 100+ keywords. Cluster: "AI video reviews" (C), "how to make AI videos" (I), "AI video pricing" (T), "AI video vs traditional" (C). Hub: "Best AI Video Tools 2025".

Example 2: "Keyword strategy for my affiliate blog about email marketing"

→ Deep keyword research. Clusters: "email marketing platforms" (C), "email automation tutorials" (I), "email marketing pricing comparison" (T), "email deliverability guides" (I).

Example 3: "Plan my content roadmap" (after monopoly-niche-finder)

→ Pick up niche from chain. Map 100+ keywords in that intersection niche. Prioritize clusters by revenue potential.

Flywheel Connections

Feeds Into

• affiliate-blog-builder (S3) — which articles to write and target keywords
• comparison-post-writer (S3) — commercial clusters become comparison articles
• content-moat-calculator (S3) — keyword count informs moat estimation
• landing-page-creator (S4) — transactional clusters become landing pages
• internal-linking-optimizer (S6) — cluster structure defines link architecture

Fed By

• monopoly-niche-finder (S1) — niche to cluster keywords for
• content-pillar-atomizer (S2) — content pillars suggest keyword areas
• seo-audit (S6) — current ranking data reveals keyword gaps

Feedback Loop

• seo-audit (S6) reveals ranking gaps in existing clusters → add keywords and new content to fill gaps

Quality Gate

Before delivering output, verify:

• Would I share this on MY personal social?
• Contains specific, surprising detail? (not generic)
• Respects reader's intelligence?
• Remarkable enough to share? (Purple Cow test)
• Irresistible offer framing? (roadmap feels actionable)

Any NO → rewrite before delivering.

References

• shared/references/seo-strategy.md — Topical authority, clustering methodology, hub-and-spoke
• shared/references/affiliate-glossary.md — Terminology
• shared/references/flywheel-connections.md — Master connection map